updated spec

zstd_compression_format.md

@@ -16,7 +16,7 @@ Distribution of this document is unlimited.
 
 ### Version
 
-0.0.2 (July 2016 - Work in progress - unfinished)
+0.1.0 (08/07/16)
 
 
 Introduction
@@ -1119,6 +1119,37 @@ with the new offset taking first spot.
   pushing the other ones by one position.
 
 
+Dictionary format
+-----------------
+
+`zstd` is compatible with "pure content" dictionaries, free of any format restriction.
+But dictionaries created by `zstd --train` follow a format, described here.
+
+__Pre-requisites__ : a dictionary has a known length,
+                     defined either by a buffer limit, or a file size.
+
+| Header | DictID | Stats | Content |
+| ------ | ------ | ----- | ------- |
+
+__Header__ : 4 bytes ID, value 0xEC30A437, Little Endian format
+
+__Dict_ID__ : 4 bytes, stored in Little Endian format.
+              DictID can be any value, except 0 (which means no DictID).
+              It's used by decoders to check they use the correct dictionary.
+
+__Stats__ : Entropy tables, following the same format as a [compressed blocks].
+            They are stored in following order :
+            Huffman tables for literals, FSE table for offset,
+            FSE table for matchLenth, and finally FSE table for litLength.
+            It's then followed by 3 offset values, populating recent offsets,
+            stored in order 4-bytes little endian each, for a total of 12 bytes.
+
+__Content__ : Where the actual dictionary content is.
+              Content depends on Dictionary size.
+
+[compressed blocks]: #compressed-block-format
+
 
 Version changes
 ---------------
+0.1.0 initial release